Activation of volume-regulated chloride currents by reduction of intracellular ionic strength in bovine endothelial cells

Abstract

1. We have studied the effects of intracellular ionic strength (Γ_i) on the swelling-activated whole-cell Cl⁻ current (I_Cl,swell) in cultured calf pulmonary artery endothelial cells (CPAE cells).

2. Reducing Γ_i from 155 to 95 mm at constant osmolarity and Cl⁻ concentration activates an outwardly rectifying current that is mainly carried by Cl⁻ ions and inactivates at positive potentials. The amplitude of the current is larger at more reduced levels of Γ_i.

3. The permeability ratio for the anions I⁻, Br⁻, Cl⁻ and gluconate (P_I: P_Br: P_Cl: P_gluc) was 1.35 : 1.03 : 1 : 0.17.

4. Blockers of the swelling-activated Cl⁻ current in CPAE cells also inhibit the current which is activated by a reduction in Γ_i with an IC₅₀ of 1.1 μM for tamoxifen, 1.3 μM for mibefradil, and 35 μM for quinidine.

5. The protein tyrosine kinase inhibitors tyrphostin B46 (50 μM) and genistein (100 μm), which inhibit I_Cl,swell in CPAE cells, also inhibited the Γ_i-induced current by 92.9 ± 2.4% (n = 3) and 41.2 ± 5.0% (n = 4), respectively.

6. Hypertonic extracellular solutions rapidly and reversibly antagonized the Γ_i-activated current, whereas increasing Γ_i from 155 to 195 mm precluded activation of I_Cl,swell by hypotonic shock.

7. It is concluded that a reduction of Γ_i activates an anion current that is identical to that activated by cell swelling. Changes in intracellular ionic strength may shift the volume set point for activation of I_Cl,swell.

Cell swelling activates in most mammalian cells a chloride current, I_Cl,swell, that is mediated by volume-sensitive anion channels (Hoffmann & Dunham, 1995; Nilius, Eggermont, Voets & Droogmans, 1996; Strange, Emma & Jackson, 1996). At the functional level, this current appears to be involved in cell volume regulation, electrogenesis, control of electrochemical gradients for ion channels and transporters, and potentially in cell proliferation and differentiation (Strange & Jackson, 1995; Nilius et al. 1996; Strange et al. 1996). The volume-sensitive anion channels are supposedly activated by a volume sensor which compares the actual cell volume with a volume set point. Under normal intra- and extracellular ionic conditions the value of the volume set point is such that a small fraction of the channels is active. Cell swelling secondary to extracellular hypotonicity progressively activates more channels. The nature of this volume-sensing mechanism as well as the identity of the volume-regulated anion channels are still unknown. It appears that a tyrosine phosphorylation step is a critical event in the swelling-induced activation of the channel since inhibitors of protein tyrosine kinases inhibit I_Cl,swell and the swelling-induced iodide efflux (Tilly, van den Berghe, Tertoolen, Edixhoven & de Jonge, 1993; Sorota, 1995; Tilly et al. 1996a; Tilly, Gaestel, Engel, Edixhoven & de Jonge, 1996b; Voets, Manolopoulos, Eggermont, Ellory, Droogmans & Nilius, 1998).

We have extensively characterized a swelling-activated Cl⁻ current (I_Cl,swell) in vascular endothelial cells of diverse origins, including calf pulmonary artery (Szücs, Buyse, Eggermont, Droogmans & Nilius, 1996) and human umbilical vein (Nilius, Oike, Zahradnik & Droogmans, 1994a; Nilius, Sehrer & Droogmans, 1994b). In these cells, I_Cl,swell is mediated by outwardly rectifying Cl⁻ channels that are inhibited by ‘classical’ anion channel blockers such as NPPB (5-nitro-2-(3-phenylpropylamino)-benzoate), NPA (N-phenylanthracilic acid) and DIDS (4,4′-isocyanatostilbene-2,2′-disulphonic acid). The channels are also inhibited by unrelated compounds, such as the anti-oestrogen drug tamoxifen, the antimalarial drugs quinine and quinidine and the Ca²⁺ channel antagonist mibefradil (Nilius et al. 1994b; Voets, Droogmans & Nilius, 1996b).

In this report we present evidence that in CPAE cells a reduction of intracellular ionic strength (Γ_i) activates a Cl⁻ current that is identical to the swelling-activated Cl⁻ current. This indicates that activation of volume-sensitive anion channels is controlled not only by cell volume but also by intracellular ionic strength.

METHODS

Materials

Tamoxifen and quinidine were purchased from Sigma. Tyrphostin B46 and genistein were obtained from Calbiochem. Dulbecco's modified Eagle's medium (DMEM), fetal calf serum, trypsin- EDTA, penicillin, streptomycin, and L-glutamine were supplied by Life Technologies (Gibco BRL). Mibefradil was a gift from Dr J. P. Clozel (Hoffmann-La Roche, Basel).

Endothelial cell culture

Calf pulmonary artery endothelial cells (CPAE) were purchased from ATCC (CCL-209) at passage 16 and used between passages 22 and 27. They were grown in DMEM containing 20% fetal calf serum, 2 mm L-glutamine, 100 units ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin, maintained at 37°C in a fully humidified atmosphere of 10% CO₂ in air, and passaged by brief exposure to 0.5 g l⁻¹ trypsin in a Ca²⁺- and Mg²⁺-free solution. Only non-confluent and non-clustered cells were used for current measurements.

Solutions

The standard extracellular solution was a Krebs solution, containing (mm): 150 NaCl, 6 KCl, 1 MgCl₂, 1.5 CaCl₂, 10 glucose, 10 Hepes, adjusted with NaOH to pH 7.4. In most experiments, we inhibited the inwardly rectifying K⁺ current by substituting Cs⁺ for K⁺ in the Krebs solution. The osmolarity of this solution, as measured with a vapour pressure osmometer (Wescor 5500, Schlag, Gladbach, Germany), was 320 ± 5 mosmol l⁻¹.

Hypotonic solutions contained (mm): 115 NaCl, 6 CsCl, 1 MgCl₂, 1.5 CaCl₂, 10 glucose, 10 Hepes, and were adjusted with NaOH to pH 7.4. This solution was made isotonic (320 ± 5 mosmol l⁻¹) by addition of 70 mm mannitol. For the 22% reduction of osmolarity, mannitol was removed.

The normal pipette solution contained (mm): 40 KCl, 100 potassium aspartate, 1 MgCl₂, 0.5 EGTA, 4 Na₂ATP, 10 Hepes, pH 7.2 with KOH (Γ= 155 mm). This solution is slightly hypotonic in comparison to the Krebs solution to avoid spontaneous activation of volume-sensitive Cl⁻ currents. Reduction of the intracellular ionic strength to 95 and 125 mm was achieved by substituting 70 or 40 mm potassium aspartate by 140 or 80 mm sucrose, respectively. In experiments with increased ionic strength we added 40 mm potassium aspartate to the normal pipette solution (Γ= 195 mm) and 80 mm mannitol to the bath solution to maintain the osmotic equilibrium. The volume-sensitive current was also under these conditions stimulated by decreasing the bath osmolarity by 22%, i.e. from 408 to 320 mosmol l⁻¹.

Current measurements

CPAE cells were used 1–3 days after seeding them on gelatine-coated coverslips (2000 cells mm⁻²). In the experimental chamber, the cells were continuously superfused with isosmotic Krebs solution at room temperature (22°C). Whole-cell membrane currents in ruptured patches were monitored with an EPC-9 (Heka Electronics, Lambrecht/Pfalz, Germany) patch clamp amplifier. The following voltage protocol was used: from a holding potential of 0 mV, a step of 0.2 s duration to −100 mV was applied, followed by a 1.3 s linear voltage ramp to +100 mV, after which the potential was stepped back to the holding potential. This protocol was repeated every 15 s.

The permeability of various anions (X⁻) relative to that of Cl⁻ (P_X/P_Cl) was calculated from the shift of the reversal potential (ΔE_rev) if extracellular chloride is partially replaced by these anions, using a modified Goldman-Hodgkin-Katz equation:

(1)

where [Cl⁻]_n and [Cl⁻]_s are the Cl⁻ concentrations in the normal and substituted external solutions, [X⁻]_s the concentration of the substituting anion, F is the Faraday constant, R the gas constant, and T absolute temperature.

Capacitance measurements

The capacitance was routinely monitored from the capacitance-track output of the EPC-9 amplifier. These values were used to normalize the current amplitudes to the cell surface area, assuming a specific membrane capacitance of 1 μF cm⁻².

Statistical analysis

Pooled data are given as means ±s.e.m. Statistical significance was calculated at the 5% level using Student's t test.

RESULTS

A reduction of the intracellular ionic strength activates a Cl⁻ current

The purpose of these experiments was to investigate the effect of intracellular ionic strength Γ_i on the volume-activated Cl⁻ current. We therefore started with some control experiments addressing the effect of Γ_i in non-stimulated cells, i.e. under isovolumetric conditions. In Fig. 1 we compare the currents measured in whole-cell patch clamp mode of a cell dialysed with a pipette solution of normal ionic strength (155 mm, left panels) and of one that is dialysed with a pipette solution of reduced ionic strength (95 mm, right panels). The current amplitude at +100 mV for each experimental condition measured during successively applied voltage ramps and normalized to the membrane capacitance is represented in the panels A and D. It is obvious that the current amplitude does not change significantly with time in the cell exposed to a pipette solution of normal ionic strength. On the other hand a large current develops during dialysis of the cell with an intracellular solution of reduced Γ_i. This current reaches a stationary level between 3 and 5 min. In panels B and E we illustrate some I-V curves, obtained at various times after breaking into the cell (marked by the filled symbols in panels A and D) and reconstructed from the currents recorded during voltage ramps.

Figure 1. Ionic currents in CPAE cells dialysed with pipette solutions of either normal (155 mm,A-C) or reduced ionic strength (95 mm, D-F).

Ionic strength was reduced by substituting potassium aspartate with sucrose, keeping the Cl⁻ concentration constant. To compare data obtained from different cells, we have expressed the current amplitudes per unit membrane capacitance. Note the different current scales for the left and right panels. A and D, time course of the current at +100 mV. Data points were obtained from the voltage ramp protocol that was applied every 15 s. At a Γ_i of 155 mm the current did not significantly change with time, whereas during cell dialysis with a solution of 95 mm Γ_i a current developed that reached a steady-state level in 3–5 min. B and E, I-V curves taken at the times marked by the filled symbols in A and D. The I-V curves a and b coincide for the cell dialysed with a solution of Γ_i= 155 mm. The main current component is an inwardly rectifying K⁺ current that is blocked by Cs⁺ (curve c). Note the presence of an outward current at positive potentials that is insensitive to Cs⁺. At reduced Γ_i, an outwardly rectifying current develops that reverses at −24 mV (curve labelled ‘diff.‘). C and F, current traces during voltage steps applied at the time indicated by the horizontal bars in panels A and D. Current traces in panel C were obtained in the presence of Cs⁺. The pattern of the currents, i.e. outward rectification and inactivation at positive potentials, is similar for both conditions but their amplitudes are very different. Voltage protocol: holding potential =−50 mV, steps between −100 and +100 mV, increment +20 mV.

For both cells, the main current component shortly after breaking into the cell (trace a) is an inwardly rectifying K⁺ current that can be blocked by replacing extracellular K⁺ by Cs⁺ (trace c in panel B). Both cells also show an outward current at positive potentials (trace a), which has been shown (Voets et al. 1996a) to represent the volume-sensitive Cl⁻ current that is partially activated under isotonic conditions. The I-V curve recorded 2–3 min after breaking into the cell (trace b) largely coincides with trace a if the cell is dialysed with a solution of normal ionic strength (panel B). The corresponding I-V curve for the cell dialysed with a solution of Γ_i= 95 mm, however, shows that at this time an outwardly rectifying current has been activated (trace labelled ‘diff.‘). The reversal potential of this current is −24 mV, which is close to the Cl⁻ equilibrium potential of −32 mV, indicating that this current is mainly carried by Cl⁻ ions. The moderate outward rectification of this current is also apparent from the current traces shown in panel F, which have been recorded during voltage steps applied after the current has reached a new steady-state level (indicated by the horizontal bar c in panel D). These traces also illustrate that the current is largely time independent, except for a slow inactivation at very positive potentials. The corresponding current traces under control conditions (panel C) show a similar inactivating pattern at positive potentials. This inactivation pattern is characteristic of the volume-activated Cl⁻ current, and suggests that the current corresponds to the volume-sensitive Cl⁻ current (Voets et al. 1996a).

The size of the current at +100 mV depends on Γ_i: it increases from 5.7 ± 0.5 pA pF⁻¹ (n = 34) at Γ_i= 155 mm to 18.7 ± 0.9 pA pF⁻¹ (n = 7) at Γ_i= 125 mm and to 56.1 ± 4.8 pA pF⁻¹ (n = 34) at Γ_i= 95 mm. Also the rate at which the current develops depends on Γ_i. The time for half-maximal activation is 120 ± 4.5 s (n = 11) at Γ_i= 95 mm and 215 ± 14.1 s (n = 7) at Γ_i= 125 mm.

Since we did not buffer calcium in the pipette solution with EGTA we were able to monitor potential changes in [Ca²⁺]_i (Ca²⁺ measurements as described in Nilius et al. 1994a). However, in none of the cells did we detect substantial changes in [Ca²⁺]_i during the course of the experiment. Also the changes in membrane capacitance never exceeded more than 4% in each tested cell (>100). Similarly, the cell volume, as monitored by the 360 nm fluorescence signal (Heinke, Raskin, De Smet, Droogmans, Van Driessche & Nilius, 1997; Nilius, Szücs, Heinke, Voets & Droogmans, 1997b), did not change significantly during the course of the experiment.

In order to differentiate between an effect of reduced ionic strength or of reduced intracellular K⁺ concentration, we replaced potassium aspartate in the normal pipette solution by the sodium, caesium, NMDG or TEA salts. None of the substitutions induced a current (data not shown). It is therefore unlikely that the reduced K⁺ concentration in the pipette solution with lowered ionic strength is responsible for activation of the current. Similarly, activation of the current cannot be attributed to changes in intracellular Cl⁻ since the pipette solutions with normal or reduced ionic strength contain identical Cl⁻ concentrations. Changes in the aspartate concentration can also be ruled out because substitution of aspartate by Cl⁻ in the pipette solution gives identical results.

The current activated by a reduction of Γ_i is identical to I_Cl,swell

The outward rectification, the inactivation at positive potentials and the reversal potential of the current activated by reduced Γ_i are reminiscent of those of the swelling-activated current. Additional evidence for the proposition that both currents are identical emerges from the comparison of their permeation properties and pharmacological profile.

The relative anion permeability of the pathway activated by reduced Γ_i was determined by replacing 150 mm extracellular Cl⁻ by an equimolar amount of iodide, bromide or gluconate (Fig. 2). Equation (1) was used to convert the measured shifts in reversal potential to permeability ratios, which are shown in the inset. The calculated anion permeability sequence P_I > P_Br≈P_Cl ≫ P_gluc is identical to that of the volume-activated anion current (Nilius et al. 1996; Strange et al. 1996).

Figure 2. Anion permeability of the current pathway activated by reduced Γ_i.

A cell was dialysed with a pipette solution of Γ_i= 95 mm and I-V curves were recorded during the plateau phase of the current. NaCl in the extracellular solution was replaced by either NaI, NaBr or sodium gluconate. The inset shows the anion permeability relative to that of Cl⁻ (P_X/P_Cl) as calculated from the observed shifts in reversal potential by means of eqn (1).

We have also evaluated the sensitivity of the Γ_i-activated Cl⁻ current for three compounds, which are known to inhibit I_Cl,swell in CPAE cells. Tamoxifen (10 μm), the antimalarial drug quinidine (100 μm) and the cardioprotective Ca²⁺ channel antagonist mibefradil (10 μm) substantially and reversibly inhibited this current as well (Fig. 3A). The pooled data from all experiments are summarized in Fig. 3C. This figure shows the mean blocking effect at +100 mV calculated from the equation:

(2)

where INH is the inhibitory effect (in%), I_max the maximal current activated by the reduced ionic strength, I_INH the minimal current after application of the inhibitor (tamoxifen, quinidine, mibefradil), and I_back the background current at +100 mV immediately after breaking into the cell. The more than 100% inhibition by tamoxifen is due to its inhibitory effect on the background current, which as shown previously represents a partially activated I_Cl,swell (Voets et al. 1996a). The concentrations of quinidine, mibefradil and tamoxifen needed for half-maximal inhibition (IC₅₀) were obtained from the fits of the dose-response relationships (Fig. 3C) to the equation:

(3)

where [inhibitor] is the concentration of the inhibitor. The IC₅₀ values tamoxifen, mibefradil, and quinidine were 1.1, 1.3 and 35 μm, respectively (Fig. 3C). The effects of these compounds are very similar, both qualitatively and quantitatively, to their inhibition of I_Cl,swell.

Figure 3. Pharmacology of the current activated by reduced Γ_i.

A, time course of the current activated at +100 mV during cell dialysis with a solution of Γ_i= 95 mm. At the plateau of the current, 10 μm tamoxifen (tam), 100 μm quinidine (quin), 10 μm mibefradil (mib) or 50 μm tyrphostin B46 (tyr) were applied. Each of these compounds produced a reversible inhibition of the current. B, inhibitory effects of genistein (100 μm) and tyrphostin B46 (50 μm). Inhibition (INH), expressed as%, was calculated by eqn (2). C, concentration-response relationships for tamoxifen (○), mibefradil (□), and quinidine (⋄). IC₅₀ values, as estimated by eqn (3) are 1.1, 1.3 and 35 μm for tamoxifen, mibefradil, and quinidine, respectively.

Figure 3A and B also shows the effect of the tyrosine kinase inhibitor tyrphostin B46 (50 μm) on the current activated by reduced Γ_i. Inhibition of the current at +100 mV is fast and nearly complete, while recovery of the current upon washout is rather slow. Genistein, another tyrosine kinase inhibitor, was less potent (Fig. 3B). Because I_Cl,swell appears to be regulated by tyrosine kinases (Tilly et al. 1993; Nilius et al. 1996; Tilly et al. 1996a; Voets et al. 1998), these findings provide additional evidence that both currents are identical.

Intracellular ionic strength modulates the volume sensitivity of I_Cl,swell

These data indicate that cell swelling and a reduction of Γ_i activate an identical anion current. We then addressed the question of whether changes in ionic strength interfere with the swelling-activated current and vice versa, and whether changes in extracellular osmolarity affect the Γ_i-induced current.

Increasing Γ_i reduces the swelling-activated current

In these experiments we have compared volume-sensitive currents activated by a hypotonic challenge of 22% in cells dialysed with the normal pipette solution (Γ_i= 155 mm) and in cells dialysed with a pipette solution of increased ionic strength (Γ_i= 195 mm). The Cl⁻ concentration was identical in both pipette solutions. Figure 4 shows an example of such an experiment. Under conditions of normal Γ_i (155 mm), the hypotonic challenge activates with a short latency a current that reaches half-maximal activation in about 90 s (Fig. 4A, current at +100 mV; for a detailed description see Nilius et al. 1996). The current is outwardly rectifying, reverses close to the theoretical Cl⁻ equilibrium potential and inactivates at positive potentials (Fig. 4B and C). The same hypotonic stimulus applied to a cell with a Γ_i of 195 mm still activates a current with similar characteristics, but its activation is clearly delayed (Fig. 4D). The current is outwardly rectifying and inactivates at positive potentials (Fig. 4E and F). The amplitude of the outward current at positive potentials is much smaller than under conditions of normal ionic strength: the current at +100 mV had an amplitude of 73.5 ± 1.8 pA pF⁻¹ (n = 13) for Γ_i= 155 mm, but amounted to only 16.4 ± 2.4 pA pF⁻¹ (n = 7) for Γ_i= 195 mm.

Figure 4. Effect of Γ_i on the volume-activated current, I_Cl,swell.

I_Cl,swell was activated by reducing the extracellular osmolarity with 22% (marked by the horizontal bar) in a cell dialysed with a pipette solution of normal ionic strength (155 mm, left panels) and in a cell where Γ_i was enhanced to 195 mm (right panels). (For more details, see Fig. 1). A and D, time course of the current activated at +100 mV. At normal Γ_i, the current activates with a short delay and reaches half-maximal activation after about 90 s. At the higher Γ_i the onset of activation was clearly delayed. The current density in the latter cell was less than 10% of that in the former cell. B and E, instantaneous current-voltage relationships at the times labelled a-e and marked by the filled symbols in panels A and D. The most prominent current component in a and e is an inwardly rectifying K⁺ current blocked by substituting extracellular KCl by CsCl (b). Challenging the cell with a 22% hypotonic solution activates a pronounced outwardly rectifying current if the cell is dialysed at normal ionic strength (c) which is largely suppressed at increased Γ_i (e). C and F, current traces recorded during voltage steps applied at the gap in the traces shown in panels A and D. Note the inactivation at positive potentials of the currents activated by hypotonicity.

Extracellular hypertonicity inhibits the effect of reduced Γ_i

The aim of this experiment was to test whether the current activated by a reduction of intracellular ionic strength is volume sensitive. The current was activated by reducing the intracellular ionic strength to 125 or 95 mm. After activation of the current, the cells were challenged with hypertonic solutions by adding either 50 or 100 mm mannitol to the extracellular bathing solution in order to shrink the cells. Figure 5A shows such an experiment for the current activated by an ionic strength of 95 mm: adding 50 or 100 mm mannitol to the isotonic bath solution caused a rapid and pronounced reduction of the current. These effects were completely reversible, although the recovery from a challenge with 100 mm mannitol was much slower. Figure 5B represents the percentage inhibition as a function of added mannitol concentration for the ionic strengths 125 and 95 mm. The current activated at Γ_i= 125 mm is completely abolished in the presence of 50 mm mannitol, whereas the current at Γ_i= 95 mm is inhibited by 70 ± 5.4% at the same mannitol concentration.

Figure 5. Effect of extracellular osmolarity on the current activated by reduced Γ_i.

A, time course of the current activated at +100 mV by a pipette solution of 95 mm ionic strength. Increasing the osmolarity of the extracellular solution by adding 50 or 100 mm mannitol clearly inhibits the current. This inhibition is also reversible, although the recovery after application of 100 mm mannitol is quite slow. B, summary of the effects of 50 and 100 mm mannitol on the currents activated by pipette solutions of 125 and 95 mm Γ_i. Data represent the percentage reduction of the maximal current observed in the isotonic solution in the absence of mannitol.

DISCUSSION

We have shown that decreasing intracellular ionic strength under isotonic conditions activates a Cl⁻ current that is identical to the swelling-activated Cl⁻ current I_Cl,swell. Our conclusion is based on the following observations. (i) The anion permeability sequence is identical for both currents: P_I > P_Br≈P_Cl ≫ P_gluc. (ii) Both currents show a modest outward rectification and inactivate at positive potentials. (iii) Compounds, such as tamoxifen, quinidine and mibefradil, block both currents in a similar concentration range. (iv) Both currents are sensitive to tyrosine kinase inhibitors. Table 1 gives a quantitative overview of the similarities between both currents.

Table 1.

Comparison of the Γ_i-activated current and I_Cl,swell

	Γi-activated current	ICl,swell	Reference
PI: PCl: PBr: Pgluc	1.35 : 1 : 1.03 : 0.17	1.37 : 1: 1.07 : 0.18	Voets et al. 1998
Quinidine (IC50)	35 μm	30 μm	Voets et al. 1996
Mibefradil (IC50)	1.3 μm	5.4 μm	Nilius, Prenen, Kamouchi, Viana, Voets & Droogmans, 1997a
Tamoxifen (IC50)	1.1 μm	3.8 μm	Voets, Szücs, Droogmans & Nilius, 1995
Genistein (100 μm)	41.2% inhibition	67.3% inhibition	Voets et al. 1998
Tyrphostin B46 (50μm)	92.9% inhibition	94.8% inhibition	Voets et al. 1998

An important question is whether the activating effect is due to a decrease in ionic strength or whether there is a specific effect of K⁺ or Cl⁻ ions on I_Cl,swell. Jackson, Churchwell, Ballatori, Boyer & Strange (1996) concluded that the volume-sensitive current in skate hepatocytes is modulated by changes in intracellular Cl⁻, since increasing intracellular Cl⁻ reduced I_Cl,swell and vice versa. However, their data do not exclude a modulation by ionic strength since in varying intracellular Cl⁻ they also changed ionic strength. On the other hand, we could clearly discriminate between changes in intracellular ionic strength and intracellular Cl⁻ concentration, since the latter was kept constant in our experiments. We could also exclude a modulation by intracellular K⁺ since a reduction of the pipette K⁺ concentration at constant ionic strength did not activate any current. We therefore conclude that a decrease in ionic strength can activate I_Cl,swell even in the absence of cell swelling. The observation that the cell-swelling-induced loss of amino acids from trout erythrocytes is enhanced by lowering the cytoplasmic ionic strength is compatible with our results, since this efflux has been shown to occur through anion channels activated by cell swelling (Motais, Guizouarn & Garcia-Romeu, 1991).

It is unlikely that activation of I_Cl,swell by reduced Γ_i proceeds via an increase in cell volume. First, the solutions with reduced ionic strength were osmotically balanced with mannitol thereby preventing osmotic gradients. Second, there was no evidence of swelling in cells that were perfused with a reduced ionic strength solution. On the other hand, cell swelling following a reduction of extracellular osmolarity may also reduce Γ_i. This prompts the question of the extent to which reduced ionic strength contributes to the activation of I_Cl,swell after cell swelling. Bulk changes in Γ_i most probably do not occur during cell swelling under ‘patch clamp’ conditions since the continuous equilibration of the cytosol with the pipette solution keeps Γ_i constant. Even a localized decrease in Γ_i near the cell membrane during cell swelling is unlikely since hypertonic pipette solutions activate I_Cl,swell in a cell exposed to normal physiological solutions (Nilius et al. 1994a,b). Therefore, the activating stimulus for I_Cl,swell under patch clamp conditions seems to be cell swelling as such, but not reduced Γ_i. In contrast, a decrease in ionic strength is expected to occur in swollen cells that are not ‘patch clamped’. Hence, both cell swelling and reduced ionic strength would contribute to the activation I_Cl,swell under these conditions.

Very recently, changes of the intracellular electrolyte composition in C6 rat glioma cells have been reported to change the set point of the activation of the volume-regulated anion current and related taurine fluxes (Emma, McManus & Strange, 1997). These data are in accordance with our findings.

How does ionic strength affect I_Cl,swell? We have recently shown that a tyrosine phosphorylation step is required for current activation by cell swelling (Voets et al. 1998). The activation by reduced ionic strength is also counteracted by inhibitors of protein tyrosine kinases, suggesting that the effect of ionic strength and cell swelling is mediated via the same tyrosine phosphorylation step. The observations that I_Cl,swell is largely inhibited by increased Γ_i and that reducing Γ_i causes spontaneous activation of I_Cl,swell at normal cell volume are therefore consistent with a model in which cell volume and ionic strength independently modulate the tyrosine phosphorylation step required for current activation.